Quick answer
Claude Code documents max 5× prompt allowance as No prompt count is published; five-hour limits were doubled on 2026-05-06. Anthropic defines Max 5× as five times the Pro per-session allowance, shared across Claude and Claude Code, and no longer publishes a prompts-per-window estimate.
Verified Sep 17, 2026Official source
Current limits
| Constraint | Current value | Scope | Verified source |
|---|---|---|---|
| Max 5× prompt allowanceAnthropic defines Max 5× as five times the Pro per-session allowance, shared across Claude and Claude Code, and no longer publishes a prompts-per-window estimate. | No prompt count is published; five-hour limits were doubled on 2026-05-06 | Max 5× | AnthropicSep 17, 2026 |
Why does this limit matter?
The broad range reflects how much code, context, tool work, and model compute each prompt requires.
This value is scoped to Max 5×; a different plan, runtime, model, endpoint, region, or account can produce a different effective constraint.
What should you check?
- Use /status before starting parallel agents and record which model is selected.
- Confirm the exact plan, model, runtime, endpoint, region, and account that serve the failing workload.
- Record the observed value, response headers or configuration, timestamp, and source without logging secrets.
Important caveats
- The earlier prompt ranges predate the May 2026 doubling of Claude Code five-hour limits and the removal of the peak-hours reduction; they are retained only as history.
- Treat the official source and live account configuration as authoritative if they differ from this verified snapshot.
HyperObserve reports the documented platform constraint. Your application, SDK, gateway, provider, region, or account can impose a lower effective limit.
Related